It is known that it is not possible to store electrical energy without an excessive outlay and therefore the electrical energy should be generated at precisely the moment at which it is consumed by the consumers. The more it is possible to determine the consumption accurately in advance, the more efficient the scheduling of the generation of the electrical energy can be and the smaller the reserve capacity should be. The longer the schedule for generating electrical energy, the easier it is to implement. Whereas it is easy to schedule the operation by way of example of coal-fired power stations for one day in advance, this is practically not possible in a short time period of by way of example a few 100 s. It is therefore desirable that loads have as few incalculable fluctuations as possible precisely in this short time period, in other words the loading graph is optimized within a short time period to as constant a value as possible.
It is also known that loads that have an energy storage capability, by way of example electro-thermal furnaces, are frequently used in industrial energy distribution networks. Energy is stored in the form of thermal energy which renders it possible within specific limits, by way of example within a cycle time period of a few 100 s, to achieve flexibility with respect to time when supplying electrical energy. A load is operated at its rated output within the cycle time period depending upon the average energy to be supplied for a control time period and is switched off for the remainder of the cycle time period. When the control time periods are suitably established within the cycle time period, it is possible to configure the summated consumption of a plurality of loads accordingly in a uniform manner. A corresponding example for a procedure of this type is described by way of example in the patent document EP 0563790 A2.
It has proven to be disadvantageous that despite optimizing the load distribution, the summated loading graph is frequently undesirably irregular over a cycle time period and that possibly a longer time period can be specified for calculating the optimization procedure.